The Hermeneutics of Computer-Generated Texts

ABSTRACT:As cultural circumstances become increasingly digital, the importance of theoretical frameworks guiding calculated considerations of authorial intention and reader response is being reaffirmed. The framework proposed in this article is that of hermeneutics: the study of understanding, of processes of meaning-making. Although explicit application of hermeneutics has fallen out of fashion, the field is especially valuable for critically approaching digital texts. This article thus serves as a reintroduction to hermeneutics, particularly for digital textual study. It offers an overview of historical hermeneutical views, and then applies a hermeneutics perspective to a new kind of text made possible by digital technologies: computer-generated prose. Through analysis and repurposing of OpenAI's GPT-2 software, this paper argues that the reintegration of hermeneutics in digital textual studies may contribute to more comprehensive understandings of both human and computer intention, especially in instances of computer-generated texts. Digital technologies are changing conventional understandings of authorship and reader responsibility; hermeneutics helps us understand what these changes are, how they have come to be, and why they matter.